A photo-electron spectrometer of hemispherical deflector type according to prior art is illustrated FIG. 1. In a photo-electron spectrometer 1 of hemispherical deflector type, a central component is the measurement region 3 in which the energies of the electrons are analysed. The measurement region 3 is formed by two concentric hemispheres 5, mounted on a base plate 7, and with an electrostatic field applied between them. The electrons enter the measurement region 3 through an entrance 8 and electrons entering the region between the hemispheres 5 with a direction close to perpendicular to the base plate 7 are deflected by the electrostatic field, and those electrons having a kinetic energy within a certain range defined by the deflecting field will reach a detector arrangement 9 after having travelled through a half circle. In a typical instrument, the electrons are transported from their source (typically a sample 11 that emits electrons after excitation with photons, electrons or other particles) to the entrance 8 of the hemispheres by an electrostatic lens system 13 comprising a plurality of lenses L1-L3 having a common and substantially straight optical axis 15.
For the following description, a Cartesian coordinate system with its z-axis along the optical axis 15 of the lens system 13 (in most cases an axis of rotational symmetry) will be used, and with the hemispheres symmetrical with respect to the (y, z) plane.
The directions of the electron trajectories are described by their angles θx against the (y, z) plane and θy against the (x, z) plane.
The lens system 13 and the detector arrangement 9 will only accept electrons which are emitted within a limited area perpendicular to the lens axis 15 and within a limited angular range. Furthermore, the source has to be positioned within a narrow range in the z-direction to achieve the best properties (in terms of sensitivity and resolution). This necessitates mounting the sample on a manipulator 17 allowing both translations and rotations in all coordinate directions, i.e. six degrees of freedom.
In many applications of for example Angle Resolved Photoelectron Spectroscopy (ARPES) a complete measurement requires full detection of a solid angle with a total cone opening of 30 degrees from a well aligned sample. Depending on sample and excitation energy/kinetic energy the required angular range may vary. The angle resolution requirements also varies with application but typically range from 1 degree down to better than 0.1 degrees. In energy resolution the desired span is from 0.5 eV down to 0.5 meV depending on application. In order to achieve a high resolution measurement the analyser arrangement must have sufficient angular and energy resolution, but since the hemispherical analyser arrangement only accepts electrons emitted within a limited angular range perpendicular to the lens axis 15, the sample manipulator 17 must have very high precision movements and repeatability. The manipulator 17 is needed to accurately rotate and tilt the sample to build up the complete 30 degree solid angle data set.
The energy distribution of electrons emitted from a sample is subject to a thermal broadening which is given by ΔE=3.5*kB[eV/K]*T[K], where ΔE is the energy distribution in eV, kB is the Boltzmann constant, and T is the temperature in Kelvin.
Therefore, in order to achieve the desired energy resolution it is imperative that the sample 11 can be cooled to very low temperatures, e. g. <1 meV broadening requires a sample temperature of at most 3 K.
The hemispheres 5 disperse the electrons with respect to their energy along the y-direction in the detector plane (which coincide with the plane of entrance 8 of the measurement region 3 in the hemispherical analyser arrangement). In the x-direction, the position in the detector plane is a direct image of the x-coordinate in the plane of the entrance 8 of the hemispheres 5. The entrance 8 of the hemispheres 5 is formed as a narrow slit in the x-direction, hereinafter referred to as the entrance slit of the measurement region or simply the entrance slit. When electrons are allowed into the hemispheres 5 through the narrow entrance slit 8, a two-dimensional detector arrangement 9 will simultaneously give information about the energy distribution and the distribution along the entrance slit 8. The two-dimensional detector arrangement 9 typically comprises a multichannel electron-multiplying plate (MCP) 19 which is arranged in the same plane as the entrance slit 8 of the hemispheres 5 and which generates a measurable electrical signal at the position of an incoming electron, which can then be registered either optically by a phosphorous screen and a video camera 21 or as an electrical pulse e. g. on a delay line or a resistive anode detector. Alternatively, some of the energy-selected electrons may be analysed further, in particular with respect to their spin, after leaving the hemisphere region through an exit aperture 23 leading to a spin detector 25. In one type of spin detector, electrons which leave the hemispheres 5 with a direction close to the (negative) z-direction are transmitted through a sequence consisting of a first lens system, a 90 degree deflector and a second lens system onto a target, after which the distribution of the scattered electrons is measured. Some instruments include two such spin detectors mounted with the deflectors at 90 deg angle to each other (i.e. one bending in the (y, z) and one parallel to the (x, z) plane), with their entrance apertures sitting in the (y, z) symmetry plane of the hemispheres, at different radial (y) positions on each side of an MCP detector.
For a given electrical field between the hemispheres 5, electrons of one particular kinetic energy, called the pass energy (Ep), will hit the centre of the MCP detector 19, and a range called the energy window will fall within the sensitive area of the MCP. The energy dispersion (dy/dE) is inversely proportional to Ep, while the energy window is directly proportional to Ep. In order to achieve a suitable compromise between energy resolution and information rate, it is thus usually necessary to adjust the kinetic energy Ek of the emitted electrons to the proper pass energy. This energy adjustment is performed by the lens system 13. This consists of a series of lens elements L1-L3 in form of concentric electrodes (cylinders, truncated cones, apertures, etc.) arranged along the optical axis 15, each connected to a voltage supply. Besides providing the energy adjustment (acceleration or retardation), the lens system 13 also allows placing the sample at a convenient distance from the hemispheres 5, and, most important in the present context, it can provide control of the distribution of the electrons in the plane of the entrance slit 8 of the hemispheres. The acceleration or retardation is controlled directly by the potential difference between the sample 11 and the hemisphere entrance 8, while the other lens voltages are used to control the electron distribution. The lens system 13 can be operated in two different modes, referred to as imaging and angle-resolving (angular) mode, respectively. In the imaging mode there is (to first order) a point-to-point correspondence between the point of emission and the (x, y) position in the plane of the entrance slit 8, independent of the take-off angle from the sample 11. The entrance slit 8 will then select electrons which are emitted from an area of the sample with the same shape as the entrance slit, and a size given by the lens magnification, i.e. normally within a narrow range in the y-direction. In the angular mode, the lens voltages are instead arranged in such a way that electrons emitted with the same angle (θx, θy) against the lens axis are focused to the same point (x, y) in the plane 26 of the entrance slit 8, as illustrated in FIG. 2, in which the y and z axes are drawn in arbitrary units and to different scales. Here, the final position is to first order independent of the start position, which is then fairly uncritical. The electrons accepted by the entrance slit 8 then have their take-off angles in the y-direction within a narrow range, defined by the entrance slit width and the angular dispersion (dy/dθy), while different take-off angles in the x-direction are distributed along the entrance slit 8. The angular dispersion is however equal in the x and y directions due to the rotational symmetry (dx/dθx=dy/dθy) of the lens system. Both the magnification in the imaging mode and the angular dispersion in the angular mode can be chosen at will and kept constant over large ranges in (Ek/Ep) by adjusting the lens voltages according to pre-calculated functions. The energy resolution of the hemispheres 5 at a given pass energy is influenced both by the width of the entrance slit 8 and the angular spread of the electron beam in the radial direction as it enters the hemispheres (i.e. spread in dy/dz). For each size of the entrance slit 8 there is a corresponding angular spread which gives the optimum combination of intensity and resolution. For narrow entrance slits, i.e. high energy resolution, the corresponding angular spread is quite small, typically 1-2 deg. This angular spread is defined by combining the entrance slit 8 with another slit 27 (hereinafter referred to as the aperture slit) some distance before it, as illustrated in FIG. 3. In the direction along the entrance slit ((x, z)-plane) there are no such restrictions in angles from resolution requirements. Since the exit angle against the median (y, z) plane after the hemisphere is the same as the entrance angle (dx/dz) against this plane (see in FIG. 1 the trajectory in the median plane and the trajectory in another plane), the directions of those electrons which are intended to reach the spin detector entrance apertures have to be quite close to the z-direction, however.
In order to compensate for misalignment of the emission point of the emitting sample 11 with respect to the optical axis 15 of the lens system 13, one deflector acting in the x-direction and one deflector acting in the y-direction are normally incorporated in the lens system. The x- and y-deflectors may be placed after each other along the lens axis 15, but more often they are combined into one deflector package 29 of four electrodes, each of which covers an azimuthal angle close to 90 deg (see FIG. 1).
Below, some of the problems with particle spectrometers according to prior art will be discussed with reference to FIG. 1. For convenience, the discussion will mainly refer to the angle-resolving (angular) operating mode of the lens system 13. It should be understood, however, that most of the arguments can be equally well applied to mapping in the imaging mode.
The requirement of efficient cooling to very low temperatures implies that the sample 11 has to be in very good thermal contact with the cooling agent, and also efficiently shielded from heat radiation. This is in conflict with the mounting on a manipulator 17 with sufficient degrees of freedom to cover the entire angular range. Mechanical movement of the sample 11 also introduces the risk that the emitting area or the area that is visible to the analyser arrangement of the spectrometer is changed, so spectra taken at different angles are also inadvertently taken from different parts of the sample.
To some extent, it is possible to avoid moving the sample 11 by using the x-deflector and/or the y-deflector described above to guide electrons starting either off-axis (in imaging mode) or with a direction which is not along the optical axis 15 of the lens system 13 (in angular mode) to the centre of the entrance slit 8. A method presented in JP58200144 A2 provides one variation on this theme. The practical applicability of any such approach is quite limited, however, since the trajectories which reach the centre of the entrance slit 8 with this technique will in the general case make an angle against the optical axis 15. For deflection in the y-direction (across the slit) they will then either be stopped by the angle-defining combination of the aperture slit 27 and the entrance slit 8, or give rise to unacceptable loss of energy resolution. In the x-direction (along the slit) only trajectories within a relatively small initial angular range will exit within the angular range accepted by a spin detector system. If the intention is to make use of the entire distribution along the entrance slit 8 there is typically the additional problem that the angular scale is severely distorted even for quite small deflections.
Furthermore, the achievable angular resolution is dependent of the angular dispersion of the lens system 13. This is most clearly seen in the y-direction, where the resolution cannot be better than the entrance slit width divided by the angular dispersion. From this point of view, it is often desirable to be able to work with a large dispersion. On the other hand, the range in θx that can be observed is limited by the smaller of (length of the hemisphere entrance slit)/(angular dispersion) and the acceptance of the lens front aperture. With increasing dispersion the limitation due to the length of the hemisphere entrance slit may then be too severe, and much smaller than the acceptance of the lens.